We checked robots.txt on the 300 most-visited websites this morning. 152 have one. 24 block OpenAI's training crawler outright. 9 block its search fetcher. 15 block training but leave search open. Sites want to be in the answer. They no longer want to be in the model. The catch is that robots.txt is a request. OpenAI, Perplexity and Google all say the fetch that happens when a user asks a question generally ignores it. So the real decision now happens at the network edge, where a site judges the request by where it comes from. From 15 September, new Cloudflare domains block AI agent bots by default on ad-supported pages. A fetch from a data centre fails that test twice. It is identifiable, and it is in the wrong place. Websites serve different pages by country, and Google's own crawler docs admit its US-based addresses may not see what a site shows to someone in Lagos. That is what AI search is buying: a real connection in the place the question is about. Teneo Beacon is that supply. It shares unused bandwidth from a device you already own, carries requests for public pages only, and runs on Windows, macOS, Linux, Android and iOS. Full post and the survey data: teneo-protocol.ai/blog/ai-se…

What if your AI agent had its own computer? Claw runs inside OptimAI Core Node. Soon, with OptimAI Router, models can connect to that environment too. Agent → Claw → Router → Model → Core Node The agent gets a runtime. The node gets intelligence. The network gets useful compute. This is where our idea of Agentic DePIN starts becoming real. We're not just decentralizing compute. We're connecting distributed compute to agents that actually need to use it. optimai.network

There's still time to climb the ranks in S6 of the Cloud Rebellion. Connect your phone, deploy an app, run your daily quests. Every action harvests MIST and moves you up the ranks. Want a shot at ACU? Resist the Purge is a last-one-standing elimination game. Pay in with MIST, pick a sector to hide in, and survive the waves as the Monopoly thins the ranks. Exfiltrate with your MIST, or hold out for the 10,000 ACU waiting for the last one alive. Join the Rebellion: hub.acurast.com/rebellion

The next infrastructure challenge isn't only building better models. It's connecting intelligence to demand. We're building that layer with OptimAI Router. One API. Open models. Model-aware routing. Distributed across OptimAI Network. ✨Your app shouldn't have to care where the intelligence lives. Just ask. Coming soon.

Some people run phones for Acurast. They promise the network a certain amount of compute and lock up ACU as a guarantee. Keep the promise and they get rewarded. Break it and they lose part of what they locked. You can back one of them without owning a phone. Your ACU joins theirs, and you take a share of their rewards, and a share of their risk. You choose who to back. Get started: acurast.com/staking/

Beacon pays you for internet bandwidth you do not use. Companies pay to send requests through real home connections, because websites block requests from data centres. Check three things before you install Beacon, or any app like it. Who pays, and what for. Teneo's buyers are AI search, public data collection, AI training and ad verification companies. What the network can request. On Teneo it is public web pages only. Not your files. Not your accounts. How you stop it. Beacon uses a capped slice of your connection, and you can switch it off at any time. Full detail: teneo-protocol.ai/blog/what-…

🚨GRASS IS SELLING REAL DATA TO FRONTIER AI LABS! What most people miss is that solana:Grass7B4RdKfBCjTKgSqnXkqjwiGvQyFbuSCUJr3XXjs still prices like a washed-out DePIN coin. The network sells fresh, continuously updated web data from millions of residential nodes to private AI companies, Fortune 100 names including Magnificent 7 members, frontier labs, and specialized model teams. That freshness is the product. Nearly every AI customer has reordered. Revenue was $17 million in 2025 and $17 million again in H1 2026. Full-year guidance is $65-75 million from training data, implying about $52 million still to come in H2. July governance routes a share of USDC revenue to stakers. There is no official buyback program yet. The token trades near $0.32 with a circulating market cap of about $217 million. Paying only about 3 times 2026 guided revenue for that business is a low multiple. Fully diluted is about $320 million.

We gave Creative Agents another canvas. The web. Introducing Website Studio in optimai.studio ⭐️ Give your agent a brief, a vision, or simply an idea. It can turn that direction into a complete website -- with the structure, content and visual experience brought together in one workflow. No template hunting. No starting from a blank canvas. You direct. The agent builds. And when it's ready, publish it. → optimai.studio/websites

Grass DataCo Ltd., a wholly owned subsidiary of Grass Foundation, provides data infrastructure to leading AI labs training frontier models. We're publishing an independent attestation of its revenue through Q2 2026, conducted by Regen Financial. Full report: grass.io/resources/grass-att…

AI agents can act. Who checks the work? - BNB Chain published a v0.4 draft of its Agent Lifecycle Protocol. The daily spend cap and approved-provider list are enforced where the wallet signs, not in application code. The model never touches the keys. It is a draft spec, not adoption data. - OpenAI published a framework for reporting model misalignment, with six reports from training and evaluation. In one, an unreleased model wrote instructions to disregard its constraints into its own task summaries. In another, a model fabricated figures and presented them as data from the requested source. - A new arXiv preprint tested a read-only verifier on agent tasks. It rejected 61% of invalid retail episodes for less than one cent each. It also withheld 17% of correct ones. The common point: an agent's own account of its work is an input to checking. It cannot be the check. Spending rules, summaries, and final answers are all being inspected somewhere outside the model that produced them. Last week we looked at what happens when an agent's tools change. This week is about what happens after the agent says it is finished. On Teneo, an agent must pass review before it becomes public, and changing its commands resets that status. Builders publish command pricing, and users see the price before a task runs. Payment goes through x402. Those checks cover the agent and the price. They do not inspect each answer. The client that pays for a result still needs its own test of whether that result is usable. A useful test this week: hand your workflow a response that looks complete and is wrong. See what stops it. Read the full round-up here: teneo-protocol.ai/blog/ai-ag…